+1 on this one.
The current state of CentOS6 is really not optimal: Lagging more than half a year behind Redhat means that there are security fixes for known vulnerabilities that are not applied to your system, and the same goes for bugfixes that may have an impact on overall stability of your systems (as in "lost man hours because of inaccessible services). I know that CentOS has promised to back-port critical security fixes from RHEL 6.1 to CentOS6 but that honestly makes me feel worse about the hole thing - in that case we would have a CentOS6 that was some unholy mix of RHEL 6.0 and RHEL 6.1, something I *really* don't want to be responsible for.
About a year and ago we had 8 RHEL subscriptions that were mainly maintained because we wanted to support OSS development - today we've replaced 30 of the CentOS installations we had back then with RHEL installations, simply because we had too many instability problems with virtualization on CentOS. Converting the servers to RHEL (not reinstalling - just changing a few configuration files and adding them to our Redhat Network) and installing the latest RHEL 6.1 kernel solved the problems in every single case - no exceptions.
Tell your CIO from someone who has been there that he will most likely be wasting more money on spent (or lost) man hours during the year than the subscription will cost you. If someone is relying on the services provided by the servers, there is a real risk that they will spend at least some time on waiting for you to get things back on track, and that is probably going to amount to real money lost pretty quickly.